JS heap OOM causing SIGSEGV during traffic spikes on production nodes

## Problem

Production nodes experience regular Harper process crashes at predictable times each day, coinciding with traffic spikes (email marketing campaign sends). The process OOMs during V8 garbage collection and terminates with SIGSEGV.

```
FATAL ERROR: NewSpace::Rebalance Allocation failed - JavaScript heap out of memory
PID 1 received SIGSEGV for address: 0x0
```

Stack: V8 incremental marking → MarkCompact → FatalProcessOutOfMemory → node::Abort()

## Pattern

- Crashes occur at the same time daily (~6:00 AM MST) on multiple cluster nodes
- Triggered by traffic spikes, not steady-state load
- Restarting Harper resolves it until the next spike
- Observed on HarperDB 4.4.15

## Investigation needed

- Whether the heap size limit should be increased via `--max-old-space-size`
- Whether there's a memory leak accumulating between restarts
- Whether HTTP worker count or request queuing can smooth peak allocation pressure

🤖 Filed by Claude on behalf of Kris.